🤖 AI Summary
To address the limited cross-subject generalizability of Parkinson’s disease (PD) detection models, this paper proposes a multi-stage robust recognition method based on hand-drawn graphics. The approach innovatively introduces a 2×2 image tiling strategy, integrates graphic-type prior classification, and hierarchically extracts both local and global features. An ensemble decision mechanism fuses discriminative information from multiple sources. This design significantly mitigates performance degradation caused by inter-subject variability: on the NewHandPD dataset, the method achieves 97.08% accuracy for seen subjects and maintains 94.91% for unseen subjects—a mere 2.17-percentage-point gap—outperforming existing state-of-the-art methods. The core contribution lies in the synergistic integration of structured tiling, task-guided feature disentanglement, and ensemble learning, thereby enhancing both cross-subject generalizability and clinical applicability.
📝 Abstract
Parkinson's disease (PD) is a neurodegenerative disease affecting about 1% of people over the age of 60, causing motor impairments that impede hand coordination activities such as writing and drawing. Many approaches have tried to support early detection of Parkinson's disease based on hand-drawn images; however, we identified two major limitations in the related works: (1) the lack of sufficient datasets, (2) the robustness when dealing with unseen patient data. In this paper, we propose a new approach to detect Parkinson's disease that consists of two stages: The first stage classifies based on their drawing type(circle, meander, spiral), and the second stage extracts the required features from the images and detects Parkinson's disease. We overcame the previous two limitations by applying a chunking strategy where we divide each image into 2x2 chunks. Each chunk is processed separately when extracting features and recognizing Parkinson's disease indicators. To make the final classification, an ensemble method is used to merge the decisions made from each chunk. Our evaluation shows that our proposed approach outperforms the top performing state-of-the-art approaches, in particular on unseen patients. On the NewHandPD dataset our approach, it achieved 97.08% accuracy for seen patients and 94.91% for unseen patients, our proposed approach maintained a gap of only 2.17 percentage points, compared to the 4.76-point drop observed in prior work.